Yao Zhang

$U$-spin symmetry is substantially broken in $D^0 \to K\pi\pi$ decays, with measured breaking parameters reaching magnitudes up to $4.45$.

A prominent $π^{0}π^{0}$ mass threshold enhancement, mirroring the $π^+π^-$ channel, is definitively observed in $ψ(3686)$ decays, challenging simple chiral descriptions.

Stored patch vectors from visual document retrieval systems can be inverted to reconstruct original pages, revealing sensitive information and enabling source page identification with high accuracy.

Distilling multi-billion-parameter multi-vector retrievers no longer requires caching terabytes of document tokens: entropic optimal transport allows a lightweight 149M text model to capture 95% of VLM teacher performance at 26x query speed using queries alone.

Disjointed cropping destroys visual context—redistributing a fixed pixel budget through continuous foveation around an MLLM's self-selected focus points beats task-trained baselines without any fine-tuning.

Qualitative examples show the PhysBrain 1.5 model's ability to produce end-effector trajectories and predict future scenes through spatially aligned RGB, depth, and robot-mask outputs.

Grounding exploration trajectories in an external fact-intent DAG prevents catastrophic context drift in autonomous agents, lifting hard CTF exploitation rates from ~25% to 75%.

Leveraging its own failure history, DiagEvo transforms self-play by dynamically guiding question generation to address specific reasoning weaknesses, leading to unprecedented accuracy gains.

Matched execution scores can hide up to 64.3 points of command-path failure, revealing a critical gap in evaluating LLM coding agents.